JP3412839B2 - Nonvolatile semiconductor memory device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a nonvolatile semiconductor memory device using a NAND type EEPROM among electrically rewritable nonvolatile semiconductor memory elements (EEPROM).

[0002]

2. Description of the Related Art A magnetic disk device is widely used at present as a secondary storage device of a computer. In recent years, an electrically rewritable nonvolatile semiconductor memory (EEPRO) has been used.
M) is reliability for its mechanical strength, low power consumption,
Taking advantage of features such as good portability and high-speed access,
It was used for applications such as replacing magnetic disks.
However, since there is a functional difference between the magnetic disk device and the EEPROM, in order to replace the conventional magnetic disk device as it is, control for filling this is required.

As one of the EEPROMs, a NAND type EEPROM capable of high integration is known. this is,
A plurality of memory cells are connected in series so that their sources and drains are shared by adjacent ones to form one unit, which is connected to a bit line. The memory cell usually has a FETMOS structure in which a charge storage layer and a control gate are stacked. The memory cell array is integrated and formed in a p-type well formed on a p-type substrate or an n-type substrate. The drain side of the NAND cell is connected to the bit line via the select gate, and the source side is also connected to the source line (reference potential wiring (FIG. 21)) via the select gate. The control gates of the memory cells are continuously connected in the row direction to form word lines. Usually, a set of memory cells connected to the same word line is called one page, and a set of pages sandwiched by a set of drain-side and source-side select gates is called one NAND block or simply one block (FIG. 22). Normally, one block is the minimum unit that can be independently erased.

The operation of the NAND type EEPROM is as follows. Data is erased simultaneously for the memory cells in one NAND block. That is, the selected NAND
All control gates of the block are set to the reference potential VSS and p
High voltage VPP (eg, 20V) on the well and n-type substrate
Is applied. As a result, in all memory cells, electrons are emitted from the floating gate to the substrate, and the threshold value shifts in the negative direction. Usually, this state is defined as the "1" state. Chip erasing is performed by putting all NAND blocks in the selected state.

The data write operation is performed in order from the memory cell located farthest from the bit line. NAND
A high voltage VPP is applied to the selected control gate in the block.
(For example, 20 V) is applied, and the intermediate potential VM (for example, 10 V) is applied to the other non-selected gates. Further, VSS or VM is given to the bit line according to the data. When VSS is applied to the bit line ("0" write), the potential is transmitted to the selected memory cell, and electron injection occurs in the floating gate. This shifts the threshold value of the selected memory cell in the positive direction. Usually, this state is defined as the "0" state. No electron injection occurs in the memory cell in which VM is applied to the bit line ("1" write), and therefore the threshold value remains unchanged and remains negative. The data read operation is performed by setting the control gate of the selected memory cell in the NAND block to VSS and the other control gates and select gates to VCC to detect whether or not a current flows in the selected memory cell.

In the NAND type EEPROM, it is necessary to write data in order from the page near the source line to the page on the drain side. The necessity will be described below with reference to FIG. "1" writing is at intermediate potential VM
(About 10V) is transferred to the drain of the selected memory cell,
Erased state without electron injection (ie negative threshold)
Keep FIG. 23 shows when the control gate 1 is in the selected state (VPP). Therefore, control gate 2 is not selected and VM
Is given. Further, VM (“1” writing) is also applied to the drain. FIG. 23A is a diagram when writing is performed from the source side, and FIG. 23B is a diagram when writing is performed from the drain side. FIG. 23
In the case of (a), since the threshold value of the cell Ma2 on the drain side is negative, the potential VM of the drain is surely set to the cell M on the source side.
is transferred to a1. However, in the case of FIG. 23B, if the "0" write operation has already been performed in the drain side cell Mb2 and it has a positive threshold value (for example, 3.5 V), "1" is written in the source side cell Mb1. When
Only the voltage obtained by subtracting the threshold voltage of the cell Mb2 from VM is transferred to the cell Mb1. Therefore, cell Mb1
Then, the potential difference between the control gate and the substrate becomes large, and erroneous writing may occur. As described above, the means for sequentially writing from the source side is important in the sense of preventing erroneous writing.

In the conventional magnetic disk device, accesses such as reading and writing of data are performed in units of sectors. Each of the storage areas formed concentrically on the magnetic recording medium is called a track.
This track is further divided into several tens of areas to form a storage unit called a sector. The capacity of one sector of a typical magnetic disk device used in a portable computer for personal use is 512 bytes.

When a computer stores data in a magnetic disk device, a block of data is called a file. It is the operating system that manages where in the storage device this file is stored. This is one of the important functions of the system (hereinafter OS).

FIG. 24 is a typical OS M used in a personal computer for portable use.
It is a figure which shows how S-DOS is using the storage area of a magnetic disk unit for file management. The boot area 61 is an area in which information necessary for starting the computer is stored, and is not related to the management of user files. In the 64 root directory area, a file management table in which the name of the file and the information accompanying it are recorded is stored. The root directory manages the data file in the same manner as other data files, but its contents may be the same as the root directory, but a subdirectory may exist hierarchically in the data area. These are functionally the same except for the existing area and the rank in the hierarchy, and are therefore generically called simply directories. The information about one file in the directory occupies a 32-byte area, and its contents are as shown in FIG. As described above, the data recording of the magnetic disk device uses the sector as a unit, but a logical unit called a cluster is used to allocate the area on the disk to the file. One cluster is determined by the type of disk so that it is composed of a power of 2 sectors. FIG. 25, 66
The start cluster number of is the cluster number assigned to the beginning of the file. In the FAT area 62 of FIG. 24, file allocation information for the data area is stored, which indicates in what order the clusters in the data area 65 are connected to form one file. 6
Reference numeral 3 is a backup of the FAT area 62. As described above, the file management of MS-DOS is performed based on the information of the FAT area and the root directory area, but the position and size of these areas are fixed and fixed for each type of disc. is there.

The access unit of the NAND type EEPROM is a page as described above. 4M bit NAND type E
Taking EPROM as an example, one page is 512 bytes,
One block consists of eight pages. Therefore, in an application in which the magnetic disk device is replaced with the NAND type EEPROM, conversion can be easily performed by associating one sector of the disk with one page of the NAND type EEPROM. However, one of the major differences that poses a problem when replacing the magnetic disk device with an EEPROM is the number of times data can be rewritten. In the case of a magnetic disk device, there is no limit on the number of times data can be rewritten on a medium, and its life is governed by mechanical factors such as damage caused by contact between the head and the magnetic recording medium. On the other hand, in the case of the EEPROM, the number of times data can be rewritten is only about 10 ⁴ to 10 ^{5 in the} current technology.

Further, as described above, the amount of "data" recorded in one sector of the magnetic disk device is 512 bytes, and this amount of data is the amount of information recorded and reproduced by the user. In addition to this, information necessary for recording / reproduction control is additionally written. One of such information is an error correction code (ECC code). This is information for detecting that there is an error in the read data and correcting the error when the user's data could not be read correctly due to the influence of media defects or noise. When writing data to
The ECC generating circuit calculates based on the recorded data and writes it together with the data on the medium.

The access unit of the NAND type EEPROM is a page, and although there are some differences, it can be regarded as equivalent to a sector of the magnetic disk device. In order to improve reliability, it is an effective measure to increase the physical capacity of one page larger than the amount of information recorded / reproduced by user data and allocate the redundant portion to ECC, as in a magnetic disk device. Is one.

In the conventional magnetic disk device, in order to use it as a storage device, initialization called format is required. Depending on the format, a track in which a series of a plurality of sectors are lined up on the circumference is formed on the magnetic recording medium, and a plurality of tracks (usually several hundreds to several hundreds) are concentrically formed. At this time, each sector
A specific initialization pattern is written in the user data area, and information necessary for recording / reproduction control is written. The ECC calculated is also written from the initialization pattern. On the other hand, in the NAND type EEPROM, as can be seen from the above description, it is not possible to overwrite the page which has been written once as it is done in the magnetic disk device. For example, in order to rewrite even one page of data, it is necessary to erase in units of blocks composed of multiple pages. Therefore, other pages in the block to be erased must be saved and written back after the erase if necessary. I have to. Therefore, it is often convenient to keep the erased page in the erased state until writing valid user data.

[0014]

In the case of a magnetic disk device, the life time guaranteed for the device is 20,000 operating hours,
If a fraction of this time is spent in rewriting data, it can be estimated that rewriting is performed about 10 ⁹ times at worst in a specific storage area such as a file management area. On the other hand, NAND EEPR
As described above, the OM can be rewritten about 10 ⁵ times, so if it is used in the same way as the magnetic disk device, in the worst case that the writing concentrates on only one block, the number of times is 10,000 times. I can't write. Since the actual data writing positions are dispersed in the storage area, the apparent number of times of rewriting increases more than that. However, if there is data that is not erased for a long period of time after being written once, this is equivalent to the narrowing of the writable area, and thus the number of times of rewriting decreases.

Further, like file management by the OS,
Generally, a file may be recorded in any free area of the data area, and its location is a directory and an FA.
It is managed in a table like T. Therefore, even if such file management is applied to the EEPROM in which the number of rewritable times is limited, it is easy to disperse the writing positions of files so that rewriting to the same location is not concentrated. However, the contents of the file management table are always rewritten when a file is written or rewritten, so the frequency of rewriting these table data is considerably higher than the frequency of writing individual data files. Moreover, since this table is the basic information for searching the file recording location, it must always be recorded in a fixed position. Therefore, when the EEPROM is applied to a file storage device, there is a problem in that rewriting concentrated on the management table area may lead to a short life.

Further, as can be seen from the above description, in the magnetic disk device, the sector before the effective user data is written can be checked by the ECC as in the sector in which the user data is written, and there is no problem. .
On the other hand, in the NAND type EEPROM that is controlled so that the erased page is kept in the erased state until the valid user data is written, the ECC is read when the page before the valid user data is written is read. Since the information in the included page is all "1", the ECC checking circuit determines that there is an error.

Normally, the operating system of the host computer does not try to read only the page in which valid data is not written, but when a plurality of pages are used as a unit of file management, already valid data is not read. It's possible to try to load it with the page where is written. In addition, when the block is erased as the page is rewritten, the means for controlling the saving of data cannot know whether or not another page in the block to be erased has already been written. In the case of the controller, erased pages must also be read into the buffer. In such a case, if the ECC check circuit operates on the data of the page in the erased state, as described above, it is erroneously recognized that an error has occurred even though the page is originally a normal page, and error processing is performed. The control is transferred, and the processing cannot proceed normally.

The present invention has been made in view of the above problems, and provides a non-volatile semiconductor memory device capable of achieving a life equivalent to that of a magnetic disk device and ensuring high reliability. The purpose is to do.

[0019]

In order to solve the above-mentioned problems, the present invention firstly provides a memory means having a plurality of storage areas divided by a capacity which is a unit of management, and a plurality of the storage means. When writing data to the area, a first management unit that manages the plurality of storage areas so that the plurality of storage areas are cyclically used in one direction according to a physical or logical arrangement, and the plurality of storage areas are recorded. Second management means for managing whether or not the data has been changed after a predetermined time point, and after the second management means is initialized when the predetermined condition is satisfied by writing the data. Is selected when the storage area in which the data has not been changed is selected, the data in the selected storage area is moved to another storage area, and when it is considered that the data in all the storage areas have been changed. That a control means for initializing the second management unit to subject matter.

Secondly, in the above-mentioned first construction, the time when the predetermined condition is satisfied is defined as the time when one cycle of the circulation controlled by the first control means is reached. To do.

Thirdly, a memory means for recording a file, a management table for managing the recorded position of the file, and a pointer configured to hierarchically indicate the recorded position of the management table on a plurality of floors. And a control means for controlling such that the recorded position of the management table is not fixed in the memory means but only the recorded position of the root pointer which is the root of the pointers of the plurality of floors is fixed. That is the summary.

Fourth, there is an error in the data from the memory means having a storage area in which the data and the error detection code corresponding to the data are recorded, and the data and the error detection code read from the memory means. Error detection means for detecting whether or not there exists, erasure detection means for detecting whether or not the storage area in which the data is recorded is in an erased state from the data read from the memory means, and the error detection Even if an error is detected in the data read by the means, when the erasure detection means simultaneously detects an erased state, the read data is treated as having no error.

[0023]

In the above structure, firstly, since the data writing is circulated in a plurality of storage areas in one direction, the writing is not concentrated in a specific storage area, and the data is evenly distributed over the plurality of storage areas. Done. Also, after writing, data that remains in the same place for a long time without being changed can be moved so that the number of times of writing compared to the entire multiple storage areas does not become imbalanced, so the apparent writing area It is prevented from narrowing and shortening the life.

Secondly, the selection of the storage area in which the written data has not been changed is performed with one cycle of writing, so that the apparent writing area is more appropriately prevented from being narrowed. .

Thirdly, the recording position on the memory means of the management table which is frequently rewritten is not fixed, and the pointer indicating the position of this management table is hierarchically recorded on a plurality of floors for the purpose of retrieval. The number of times of rewriting the pointer that must be recorded at a fixed position is suppressed to the same as that of other areas, and it is possible to prevent writing from concentrating on a specific storage area and reaching the end of its life in a short time.

Fourth, it is detected from the read data whether or not the page or the like, which is the storage area, is in the erased state, and it is possible to prevent an error in the detection result of the error detecting means. As a result, even when the control is performed such that the erased page or the like is read out without being distinguished from the data-written page or the like, accurate error detection of the read data is performed and high reliability is ensured.

[0027]

Embodiments of the present invention will be described below with reference to the drawings. This embodiment is applied to a nonvolatile semiconductor memory device using a NAND type EEPROM.

1 to 9 are views showing a first embodiment of the present invention. First, FIG. 2 shows an EEP as a memory means.
The structure of the storage area in ROM is shown. 3 in the figure
Reference numerals 1 to 36 denote a plurality of storage areas divided by a capacity serving as a unit of management. If the storage area is the same as the erasing unit or an integral multiple thereof, the handling becomes easy, but the present invention is not limited to this. Further, in the present embodiment, this unit storage area is simply referred to as a block hereinafter, but this may be interpreted as a block which is an erase unit of the NAND type EEPROM, but is not limited to this. The numbers given to the blocks are for ranking the blocks, and it is easy to understand that the numbers are in the physical address order, but the logical order may follow a specific rule. For the sake of simplicity, the memory means used in this description will be composed of 6 blocks.

FIG. 1 is a table for managing writing to the blocks 31 to 36. Each item of 11 to 16 indicates the state of blocks 31 to 36.
-23, it is composed of three flags. That is, a flag (hereinafter, referred to as a third management means for managing whether or not significant data is recorded in the block)
(Valid flag or V flag) 23 and the valid flag, and first management means for managing data writing so that the data writing is used in one direction in accordance with the order of blocks and in a circular manner. 21 (hereinafter, erase flag or E flag) and a flag (hereinafter, change flag or C) as second management means for managing whether or not the data recorded in the block has been changed after a certain time. Flag) 22.

Next, using the drawings and the flow chart,
The writing method of this embodiment will be described. Note that in order to read the written data, a means for managing the correspondence between the tag (for example, file name) of the written data and the written position is necessary, but this is performed by a normal file system (OS). Since the method described above can be applied, description thereof will be omitted in this embodiment.

FIGS. 3 and 4 are flowcharts showing a data writing procedure, FIG. 5 is a flowchart showing a data erasing procedure, FIG. 6 is a flowchart showing a data rewriting procedure, and FIG. 7 is a subroutine used in the writing procedure. It is a flowchart of. 8 and 9 are diagrams for simultaneously explaining changes in the states of the storage area shown in FIG. 2 and the table for managing it shown in FIG. 1 due to the writing of data. The shaded blocks in the drawings showing the storage areas in FIGS. 8 and 9 indicate areas in which valid data is written.

First, before the memory device is used, FIG.
It is assumed that all flags are initialized to "0" as shown in (a). Consider writing the first block in this state. In step 201 of FIG. 3, it is checked that the required capacity is free. Next, in this method, since both the V flag that manages whether significant data is recorded and the E flag that manages the cyclic use in one direction are written only to the block of "0". A block search that meets this condition is performed (steps 202 and 203 in FIG. 3, steps 301 to 304 in FIG. 7). After it is confirmed that the conditions are met, the data is written in the first block (steps 204 to 206 in FIG. 3), and the C flag and the V flag are set to “1” (steps 207 and 8 in FIG. 3).
(B)). If you rewrite the data you just wrote,
In accordance with the procedure of 1., V flag “0” of the first block is set to E
The flag is changed to "1", and the data is written in the second block searched by the same procedure as the first writing (see FIG. 8).
(C)). Although the first block is now in a logically erased state, it must be electrically erased separately. Although it is preferable to perform the process in the background from the viewpoint of processing speed, the method is out of the essence of this embodiment, and thus the description thereof is omitted. The following description assumes that these blocks have been electrically erased in some way before the data is written.

When the writing process is continued, the data is uniformly written in the entire memory in the order of block numbers,
For example, the state shown in FIG. If you try to write data for one block,
Although there is a free area, the search in step 203 of FIG. 3 fails. The state where the search has failed indicates that the uniform writing to the free area of the block has completed. Therefore, the E flag for all blocks is cleared to "0" to control the next unidirectional writing (step 209 in FIG. 3, FIG. 8E).

What characterizes this embodiment together with uniform writing by unidirectional circulation is that data that stays in the same place for a long time is moved. In order to realize this, a timing related to the number of times of writing, which is appropriate for performing the movement, is required, and one of them is the time when the E flag is cleared. In FIG. 3, the data to be moved is selected by inspecting the C flag in the processing of steps 210 to 213 at this point. Figure 8
In (e), the C flags of all the blocks are "1". This means that all the blocks have the lowest value from the time when the C flags are initialized to the state of FIG. 8 (e). 1
The times indicate that they have been rewritten. In this case, since it cannot be determined which data occupies the same block for a longer time, it is reinitialized in step 214 (FIG. 8 (f)). After that, the data writing of the first block which started this series of processing is performed (step 2 in FIG. 4).
24-228, 206, FIG. 9 (a).

Further, it is assumed that the data of the fifth block is erased to be in the state of FIG. 9B, and the writing of the second block and the erasing of the first block are performed to result in the state of FIG. 9C. Consider the case where the first block is written at this point as the end of the write pattern. Since the rewriting has completed another round, the search for the writable area has failed,
All E flags are cleared (FIG. 9 (d)). Then C
The flag is checked, but this time, C in the second block
Since the flag is "0", it is determined that the data in this block has not been rewritten for a relatively long period of time, and the data is moved. According to the procedure of steps 215 to 220 of FIG. 3, the write-destination block of the move destination is circulated in the writing direction starting from the block containing the data to be moved and searched. In this example, since the fourth block corresponds, the second block is erased after the data is copied there (steps 221 to 223 in FIG. 4, FIG. 9).
(E)). As a result, the data occupying the second block all the way from the time of FIG. 8C is moved, and this block becomes rewritable.

After that, it is sufficient to write the data for one block requested at the time of FIG. 9 (c) to the empty block, but even if the writable block is searched for like the normal writing, it fails. There are cases. That is, FIG.
When there is exactly one empty block at the time of (c), the moving target data enters this empty block, and the E of the block originally containing the moving target data.
Since the flag is "1", it will not be caught in the search. Therefore, the processing for clearing the E flag as shown in step 227 of FIG. 4 is necessary. If you do the control to keep the E flag of the block originally containing the data to be moved as "0", such a thing will disappear.
Instead, even though there are other free blocks, the block immediately after the movement may be searched as a rewritable block, and electrical erasing of the background may not be in time.

Next, FIGS. 10 to 14 show a second embodiment of the present invention. In this embodiment, it is assumed that a 4M bit NAND type EEPROM is used as the nonvolatile semiconductor memory. Further, in order to briefly explain the main points of the embodiment, it is assumed that the unit of logical management of the storage area corresponding to the cluster of MS-DOS matches the block which is the erase unit of the NAND type EEPROM. Further, in this embodiment, a user file, a table for managing the recording position of the file, and a hierarchical pointer to the table whose position is not fixed need to be recorded in a specific area in the storage area. Since it is not possible, it should be controlled so that writing is distributed over the entire storage area as much as possible. However, since the present embodiment is for explaining the method in which the table for managing the position of the file is not fixed to a specific area in the storage area, it is premised that distributed write control is performed. I won't go into the details.

FIG. 11 is an example of a block that constitutes one of the tables for managing the recorded position of the file, and here it is tentatively called a route management table. Reference numeral 38 in FIG. 11 is a table for managing the block allocation status of the entire storage area with the same configuration as the FAT of MS-DOS. In the present embodiment, not only the user file but also the file management table itself and its pointer Since the position of the block in which the management information is recorded also moves, their allocation is also managed. Figure 1
39 of 1 is the position of the additional block used when the allocation information of the block is not stored in this block, or M
It is assumed that the file name corresponds to the directory of the S-DOS and a pointer indicating the position of the management table or the like regarding the information associated with the file name. That is, if the management table shown in FIG. 11 can be accessed, all information related to file management can be accessed.

FIG. 12 is an example of a block in which a pointer configured to hierarchically indicate the recorded position is recorded in order to indicate the recorded position in the route management table. In FIG. 12A, 41 to 48 are NAND type EEPRs.
In the OM page, 48 is the source side and 41 is the drain side. It is assumed that the pointer has an amount of information that can be stored in one page, and is written in order from the source side according to the writing rules of the NAND type EEPROM. 12
The hatched pages 46 to 48 in (a) indicate that the updated pointer has been written.
"ULL" indicates that the page remains erased.
The block containing the first pointer in FIG.
The information on the page indicates the position of the route management table. The block storing the m-th pointer in FIG.
It is assumed that the position of the (m-1) th pointer is indicated with the same configuration. When the number of pointers (the number of layers) is n, the first to (n-1) th pointers may be recorded anywhere in the storable area, and are accessed by sequentially manipulating pointers on multiple floors. It Only the nth pointer (hereinafter, the root pointer) is recorded at a fixed position in the storage area.
FIG. 10 shows the connection of pointers set on the plurality of floors.

FIG. 13 is a flow chart showing the procedure of the processing that occurs with the update of the route management table necessary for the file management method according to this embodiment. It is assumed that the contents of the route management table have been read and expanded in advance in the RAM, and that the update is written back in the RAM. First, when the block allocation is changed by writing a file or the like, the position of the block to which the changed route management table is written back is determined by referring to its own block allocation table (step 401). Next, this position is set as a pointer, the allocation table is updated, and then writing is performed (steps 402 to 405). Steps 501 to 510 of the flowchart of FIG. 14 show a procedure for setting a block position in a hierarchical pointer. This procedure is called recursively to change the settings of layered pointers.

Next, how to determine the size of the number of hierarchical layers n of the pointer will be specifically described. Whenever the block allocation is changed by writing a file, the contents of the route management table are updated. The updated table itself is also written to another block so that writes are not concentrated on a particular block. Since the position of the route management table has changed, the recording position of the first pointer is also updated. Since the pointer can be updated 8 times (for 8 pages) in the same block, the block storing the first pointer is written once in 8 changes of the block by writing the file, and the 9th time is executed. The position changes with the update. Similarly, it can be estimated that the block storing the second pointer will be rewritten once every 64th writing of the file. Here, if the total capacity of the non-volatile semiconductor memory device is 20 Mbytes, the total number of blocks is 5,120 because one block has a capacity of 4 kbytes. If the number of rewritable times of the EEPROM is set to C times and writing is uniformly performed in these blocks, rewriting of a total of 5120 × C blocks may occur. When one block of the management table is rewritten for one block of the file, two blocks are rewritten at one time, so the maximum number of block allocation changes is 5120 × C / 2 = 2560 × C. Times.

Rewriting of the fixed block in which the n-th pointer is recorded occurs once every 8 ⁿ block allocation changes. Also, since this block can be rewritten up to C times, if the block allocation is changed up to 8 ⁿ × C = 2 ³ⁿ × C times, the number of rewriting of this fixed block reaches the limit earlier than other blocks. There is no such thing. Therefore, the required number of layers n is 2 ³ⁿ × C = 2560 × C n = log ₂ 2560 / 3≈4.

15 to 20 show a third embodiment of the present invention. FIG. 15 is a block diagram showing the overall configuration of the nonvolatile semiconductor memory device according to this embodiment. In the figure, 115 is a NAND type EEPROM as a memory means.
The M module is composed of a memory cell array divided into blocks each including a plurality of pages. EE
The PROM module 115 is connected to a host system (not shown) via the host interface 101 connected by a data line. A multiplexer 113 and a data buffer 107 are provided on the data line.
Further, in the host interface 101, a data register 102, an address register 103, a count register 104, a command register 105, a status register 106 and an error register 116 are provided.
108 is a control logic, 109 is an error control logic, 114 is an address generator, 110 is a CPU having a function as a control unit, and 111 is a work RA.
M and 112 are control program ROMs. The control program ROM 112 stores a series of control programs for writing data and the like.

FIG. 16 is a diagram showing the internal structure of the error control logic 109. ECC check circuit 150 as error detection means, E as error detection / correction code generation means
A CC generation circuit 151 and an erase detection circuit 152 as erase detection means are provided.

The memory device of the present embodiment uses a table for managing the usage status of the memory area as needed for the data recorded in the EEPROM module 115 which is a non-volatile memory area. Although this table is recorded in the EEPROM module 115 together with other user data, it is automatically read into the working RAM 111 when the apparatus is started up. Also, this table is
E each time it is updated or when use of the device ends
It will be written back to the EPROM module 115.

Next, the operation of this apparatus will be described with reference to the flow chart. The host system sets the access start address in the address register 103 in the host interface 101 of FIG. 15, sets the sector length of the data to be accessed in the count register 104, and finally sets the read / write command in the command register 105. . When an access command is written in the command register 105 of the host interface 101, the CPU 110 in the controller reads the command in the command register 105 and executes a series of control programs for command execution stored in the control program ROM 112.

FIG. 17 shows the EEPROM module 115.
4 is a flowchart showing a procedure for reading data from the. First, the CPU 110 of FIG. 15 determines the physical address of the EEPROM module 115 to be read from the start address set in the host interface 101 (step 601). Next, EEP
Data is read from the ROM module 115 to the data buffer 107 (step 602). Next, error processing and data transfer from the data buffer 107 to the host system are executed (steps 603 to 605).

18 and 19 are flowcharts showing the procedure for reading data from the EEPROM module 115 to the data buffer 107. CPU 110
EEPROM module 115 to multiplexer 113
Through the data buffer 107 and the data buffer 107 is set to the read mode (step 7).
01, 702). A physical address of the EEPROM module 115 to be read is set in the address generator 114 (step 703). And
In the data buffer 107, the area in which the read data is to be stored is determined, and the head address is determined in the data buffer 107.
Set as a write address for writing (step 70)
4). Then, the control logic 108 is instructed to execute a predetermined sequence for reading data.

The control logic 108 sets the multiplexer 113 so that the read data from the EEPROM module 115 flows to the data buffer 107, and reads the data for one sector while incrementing the content of the address generator 114 (step 705). . Further, the ECC checking circuit 150 of FIG. 16 is controlled so as to detect an error using these data and the ECC code read together with the data, and at the same time,
The erase detection circuit 152 is controlled to detect whether the read data is in the erased state. The erasure detection circuit 152, when the data for one sector and the ECC code are all "1" which is the state after erasure, the CPU 11
A code indicating that the erased state is detected is set in a register accessible from 0. When the data for one sector is read, the CPU 110 causes the ECC checking circuit 15
Check 0 for error in data (step 70)
6). If no error is detected, the data is transferred from the data buffer to the host system. When an error is detected, the CPU 110 next accesses the erase detection circuit 152, and if the erased state is detected, the result of the ECC check circuit 150 is regarded as a false detection and the data in the buffer is transferred to the host system. . If the ECC check circuit 150 detects an error and the erase detection circuit 152 does not detect an erased state, and if the detected error is correctable,
The data buffer 107 is accessed to correct erroneous data, and then the data is transferred to the host system.

If an uncorrectable error is detected, data transfer to the host system is not performed,
The CPU 110 sets a code indicating that an error has occurred in the status register 106 in the host interface 101 and a code indicating the content of the error in the error register 116, and notifies the host system that the execution of the instruction has terminated abnormally. And ends the processing (step 707-).
712).

FIG. 20 is a flow chart showing the procedure for transferring data from the data buffer to the host system. The CPU 110 sets the head address of the area in which the data read in the data buffer 107 is stored as the read address from the buffer (steps 801 and 8).
02), instructs the control logic 108 to transfer data for one sector to the host system. The control logic 108 controls the data buffer 107 and the host interface 101 to transfer data for one sector to the host system (step 803), and when this is completed, the address register 10
3 is advanced by 1 sector, 1 is subtracted from the count register 104, and the CPU 108 is notified that the transfer is completed.
As long as there is data left to be transferred to the host system,
The CPU 110 repeats this control. When all the read data has been transferred, the CPU 110 sets a code indicating that there is no error in the status register 106 in the host interface 101, notifies the host system that the execution of the instruction has ended, and ends the processing. .

[0052]

As described above, according to the present invention,
Firstly, since the data writing is performed in a single direction in a circulating manner in a plurality of storage areas, it is possible to uniformly write in the plurality of storage areas without concentrating on a specific storage area. it can. In addition, since the data that remains in the same place for a long time without being changed after being written is moved so as not to cause an imbalance in the number of rewrites compared with the entire multiple storage areas, an apparent write is performed. It is possible to prevent the area from being narrowed and the life to be shortened. Therefore, even though there is a limit to the number of times of rewriting the same storage location, it is possible to realize a life equivalent to that of a magnetic disk device and expand the application.

Secondly, the storage area in which the written data has not been changed is selected in one writing cycle, so that the apparent writing area is more appropriately prevented from being narrowed. can do.

Thirdly, the arrangement of the management table, which is frequently rewritten, in the storage area is not fixed, and the pointer indicating the position of this management table is hierarchically recorded on a plurality of floors, so that it can be searched for. It is possible to prevent the number of times of rewriting the pointer, which has to be recorded at a fixed position, from being suppressed to the same as that of other areas, and to prevent writing from concentrating in a specific storage area and reaching the end of life in a short time.
Therefore, similar to the first aspect of the present invention, it is possible to realize a life equivalent to that of the magnetic disk device, although the number of times of rewriting the same storage location is limited.

Fourthly, it is detected from the read data whether or not the page or the like which is its storage area is in the erased state,
Even if an error is detected in the read data, if an erased state is detected at the same time, the read data is processed as if there is no error, so the erased page is distinguished from the page in which the data is written. Even when the control is performed such that the data is read out without any error, it is possible to accurately detect an error in the read data and ensure high reliability.

[Brief description of drawings]

FIG. 1 is a first non-volatile semiconductor memory device according to the present invention.
It is a figure which shows the structure of the table for managing the writing to a block in an Example.

FIG. 2 is a diagram showing a configuration of a storage area in an EEPROM in the first embodiment.

FIG. 3 is a flowchart showing a procedure of writing data to a storage area in the first embodiment.

FIG. 4 is a flowchart showing a procedure of writing data to a storage area in the first embodiment.

FIG. 5 is a flowchart showing a procedure for erasing data in the first embodiment.

FIG. 6 is a flowchart showing a procedure of rewriting data in the first embodiment.

FIG. 7 is a flowchart of a subroutine used in the data writing procedure of FIG.

FIG. 8 is a diagram for explaining a change in the state of a storage area and a table for managing the storage area by writing data in the first embodiment.

FIG. 9 is a diagram for explaining a change in the state of a storage area and a table for managing the storage area by writing data in the first embodiment.

FIG. 10 is a diagram showing a connection relationship between a pointer and a route management table in the second embodiment of the present invention.

FIG. 11 is a diagram showing a configuration of a route management table in the second embodiment.

FIG. 12 is a diagram showing blocks in which pointers are recorded in the second embodiment.

FIG. 13 is a flowchart showing a procedure of processing that occurs in association with the update of the route management table in the second embodiment.

FIG. 14 is a flowchart showing a procedure for setting a block position to a pointer in the second embodiment.

FIG. 15 is a block diagram showing a configuration of a nonvolatile semiconductor memory device according to a third example of the present invention.

16 is a diagram showing an internal configuration of an error control logic in FIG.

FIG. 17 is a flowchart showing a procedure of reading data from the EEPROM module in the third embodiment.

FIG. 18 is a flowchart showing a procedure of reading data from an EEPROM module to a data buffer in the third embodiment.

FIG. 19 is a flowchart showing a procedure of reading data from an EEPROM module to a data buffer in the third embodiment.

20 is a flowchart showing a procedure for transferring the read data from the data buffer in FIG. 17 to the host system.

FIG. 21 is an equivalent circuit diagram showing one NAND cell of the EEPROM.

FIG. 22 is an equivalent circuit diagram showing a memory cell array of an EEPROM.

FIG. 23 is a diagram for explaining the write operation of the NAND type EEPROM.

FIG. 24: MS which is one of operating systems
FIG. 6 is a diagram illustrating a file management method of DOS.

FIG. 25 is a diagram for explaining a configuration example of a directory.

[Explanation of symbols]

21 A first management means, which is a flag for referring to the valid flag so as to control the data writing so that the data is written in one direction according to the order of blocks and is used in a circulating manner. A flag 23 for managing whether or not the data recorded in the block has been changed after a certain time, and a flag 31-36 for managing whether or not significant data is recorded in the block. Tables 41 to 48, 51 to 58 for managing allocation of a plurality of storage areas 38 blocks divided by different capacities Page 109 on which pointer information is recorded Error control logic 110 CPU (control means) 115 EEPROM module (memory means) 150 ECC check circuit (error detection means) 151 ECC generation circuit 152 Erase detection circuit (erase detection means)

─────────────────────────────────────────────────── --- Continuation of the front page (56) References JP 62-283496 (JP, A) JP 4-165547 (JP, A) JP 52-22834 (JP, A) JP 56- 29899 (JP, A) JP-A-4-163966 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G06F 12/00-12/06 G06F 12/16 G11C 16/02

Claims (4)

(57) [Claims] 1. A memory means having a plurality of storage areas divided by a capacity serving as a unit of management, and a memory means for writing data to the plurality of storage areas .
A first management unit that manages a plurality of storage areas so that they are circulated and used in one direction according to a physical or logical arrangement ; and when the data recorded in the plurality of storage areas is at a predetermined time.
The second management means for managing whether or not it has been changed after the point, and the predetermined condition has been established by writing the data.
At a time after the second management means is initialized
Select the storage area whose data has not been changed,
Move the data in the selected storage area to another storage area.
And the data in all storage areas are considered to have changed
And a control means for initializing the second management means. 2. A time point when the predetermined condition is satisfied
2. The non-volatile semiconductor memory device according to claim 1, wherein is defined as a point of time of a cycle of circulation managed by the first management means. 3. The memory means is a file, the file
Management table for managing the recorded position of the management table and the recorded position of the management table in a hierarchical manner on a plurality of floors.
Configured pointer is recorded on, the management table
The nonvolatile semiconductor memory according to claim 1, the position to be recorded, characterized in that only recorded the position of the root pointer to be the source of the pointer of the multi-storey not fixed is controlled to be secured apparatus. 4. The memory means stores data and the data.
Equipped with a storage area where the error detection code corresponding to
, The data read from the memory means and error detection
Error detection means for detecting whether or not there is an error in the data from the output code, and the data read from the data read from the memory means.
Further , an erase detecting means for detecting whether or not the recorded storage area is in an erased state is provided , and even when an error is detected in the data read by the error detecting means, the erase detecting means simultaneously detects the erased state. The nonvolatile semiconductor memory device according to claim 1, wherein when detected, the read data is processed as having no error.